This preview shows pages 1–12. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: The simplest neuralnetwork model for brain computations is feedforward with one output. The simplest model for the postsynaptic current in a linear feedforward network is dI s dt = I s s + w b u b t ( ) s b = 1 N s dI s dt = I s + w u We assume a steadystate currentto actionpotentialfrequency function (the activation function), F(I s ). An extreme model uses very fast firing: For very slow firing: s dI s dt = I s + w u with v = F I s ( ) r dv dt = v + F I s t ( ) ( ) r dv dt = v + F w u ( ) Neurons can display both slow and fastfiring properties as the mean input current varies. The simplest neuralnetwork model for brain computations is feedforward with one output. A full feedforward network has vector inputs and outputs connected by a weight matrix. For a feedforward network: r dv a dt = v a + F W ab b = 1 N a u b r d v dt = v + F W u ( ) Reaching with hands is independent of gaze; how does the brain transform coordinates? Premotor cortex neurons encode the site of objects in bodybased, not retinal, coordinates. objects in bodybased, not retinal, coordinates....
View
Full
Document
This note was uploaded on 06/08/2009 for the course BME 575L taught by Professor Grzywacz during the Spring '09 term at USC.
 Spring '09
 Grzywacz

Click to edit the document details